• CHRONIC HEPATITIS
• HEPATITIS B
• HEPATOCELLULAR CARCINOMA

Although safe and effective vaccines against HBV (hepatitis B virus) are available, there are worldwide more than 2 billion people who had an HBV infection and about 250 million people suffering from chronic HBV infection. Chronic HBV infection is a major cause for liver diseases such as fibrosis, cirrhosis and hepatocellular carcinoma (HCC). It is estimated that about 800 000–1 000 000 people die each year due to the consequences of chronic HBV infection.1 Moreover, in almost all HBV-associated HCCs integrated HBV-DNA is found. Therapy options at present are limited and based on nucleoside/nucleotide analogues and interferon alpha. Since persistence of HBV infection frequently can be attributed to an insufficient cellular immune response approaches to rescue host immune response may help to eliminate infected cells and to suppress virus replication. A recent development are HBV-specific CARs (chimeric antigen receptors) human T-cells that are intended to recognise and eliminate HBV positive cells.2 CARs are designed molecules consisting of several components: an extracellular antigen binding domain, frequently encompassing the scFv (single chain variable fragment) formed by the variable domains of the heavy and light chain of an antibody, a spacer and transmembrane region ensuring the anchoring of the receptor on the cell surface with certain flexibility, and an intracellular effector/signalling module frequently derived from CD3 zeta and costimulatory domains (frequently derived from CD28) mediating T-cell activation in response to specific binding of the antigen. The subsequent T-cell activation is intended to selectively eliminate the target cell. A relevant aspect of this approach is the capacity to allow killing of the target cells independent of major histocompatibility complex, which is required for target cell king by natural killer T cells. This approach has shown promising results for the therapy of leukaemias (B-cell acute lymphatic leukaemia), non-Hodgkin’s lymphoma and chronic lymphatic leukaemia) and some CAR-T cell products have gained market authorisation in the EU and USA.3 CAR-T cell-based therapies for chronic HBV are still in the early phase of development but are considered as a novel and promising therapeutic approach.

The functionality and effectiveness of CAR-based approaches strongly depend on the scFv part of the chimeric molecule. The main function of CARs so far is seen in killing target cells when expressed in T-cells or in NK cells. In their recent study entitled ‘Chimeric antigen receptors of hepatitis B virus envelope proteins inhibit hepatitis B surface antigen secretion’, Wang et al 4 extend the spectrum of CAR activity. In previous studies from the authors and from others there was evidence that application of HBV-specific Ig, in addition to its neutralising activity, exerts an inhibitory effect on the release of viral and subviral particles. This inhibitory effect was found to depend on the internalisation of the antibodies and subsequent interference with (sub)viral morphogenesis and release.5

With respect to HBV morphogenesis and release, there are still some aspects not fully understood. In HBV replicating cells, in addition to the formation of the infectious viral particles, non-infectious subviral-particles are formed in excess.6 The prominent forms of subviral particles are the small (22 nm diameter) spheres and filaments which are exclusively formed by the surface proteins of HBV (HBsAg) and lack any genome or capsid. The forms differ with respect to their composition of the HBV surface proteins. In contrast to the spheres which are mainly formed by small surface protein the filaments, as well as the virions, contain in addition a significant fraction of LHBs (large surface protein). Both spheres and filaments also contain some (MHBs) middle surface protein. Apart from the different morphology, spheres and filaments also differ with respect to their release pathways. While infectious viral particles (complete virions) and filaments are released via mulivesicular bodies (MVBs) in an endosomal sorting complexes required for transport (ESCRT)-dependent manner, the secretion of spheres occurs via the classic secretory pathway.7 8 Independent from this the initial steps of (sub)viral morphogenesis start at the endoplasmic reticulum (ER) with the synthesis of the three surface proteins which are integral membrane proteins.

In order to efficiently interfere with HBV morphogenesis and release, the authors identified two well characterised HBV-specific monoclonal antibodies (mAbs). mAb MA18/7 binds to the N-terminal domain of LHBs and thereby exerts a neutralising capacity due to interference with LHBs binding to the high affinity receptor NTCP. mAb G12 was developed and characterised in the authors’ own lab and binds to a strictly conformation-dependent epitope on the antigenic loop within the S-domain of the HBV surface proteins.9 The S-domain is common to all three HBV surface proteins. The antigenic loop is stabilised by the formation of a complex array of disulfide bonds.

Based on these mAbs, scFvs were established. To enable interference with virus morphogenesis and release it was required to target the scFvs to the ER lumen (a central compartment for the initial steps of virus morphogenesis) or to membranes as HBV surface proteins are synthesised as integral membrane proteins. Based on these considerations different constructs were designed enabling the overproduction scFv fusion proteins either with a N-terminal signal sequence (SSP) derived from IL-2 to ensure ER luminal localisation and subsequent entry to the secretory release pathways or to produce membrane associated scFvs as CARs. In contrast to the expression of scFv in HBV producing cells, the expression of the SSP-scFv fusion protein had a strong inhibitory effect on the release of HBsAg (both subviral particles and viral particles). Although the amount of recombinant protein was decreased in case of the scFv-CAR fusion protein as compared with the SSP-scFv fusion protein the inhibitory effect on the release of HBsAg was even more pronounced. A further interesting aspect is that the effect on the release of infectious viral particles is more pronounced than the effect on subviral HBsAg release. The inhibitory effect depends on the direct interaction of scFv domain with the viral surface proteins, preventing their proper assembly and release. Interestingly, the robust inhibition of the viral and subviral particle release is associated with only a minor increase of the intracellular HBsAg level. This suggests that intracellular accumulation of HBsAg is ameliorated, which would lead to ER-overload, frequently associated with increased ROS formation. This would be detrimental for a potential therapeutic use. It can be speculated that the HBsAg (virion)-SSP-scFv or HBsAg (virion)-scFv-CAR complexes are prone for rapid intracellular degradation. This also helps to reduce the risk of accumulation of HBV DNA replicative intermediates in the cell, which could lead to an elevated integration of HBV DNA or enhanced levels of the viral nuclear episome, cccDNA (covalently closed circular DNA) that is derived from the viral replicative DNA and serves as the viral genome reservoir sustaining HBV persistence. An additional interesting aspect is that the release of infectious viral particles is even more strongly affected than the release of HBsAg mostly representing spheres. As the viral particle (42 nm diameter) is bigger than the spheres there are more binding sites for the scFv-antigen receptor available. Also, in the case of the LHBs-specific MA18/7 based approaches this might reflect the fact that the LHBs content in virions is significantly higher than in spheres, which contain only very small amount of LHBs. In addition, it should be noted that spheres on the one hand and filaments and virions on the other hand leave the cell via different routes, the classic secretory pathway vs the ESCRT-MVB-dependent pathway respectively. In light of the open questions concerning HBV morphogenesis and release, HBsAg-specific scFv-CARs also represent a powerful research tool for a deeper analysis of the HBV life cycle.

In addition to the in vitro data, the authors show in vivo data based on murine HBV models, including hydrodynamic injection or adeno-associated virus transduction to deliver the HBV genome to the mouse liver. Again, significant inhibition of HBsAg and virus release can be observed in vivo underscoring the potential of this approach for further development as therapeutic strategies. With respect to potential therapeutic applications it is reassuring to see that there is no obvious intracellular accumulation of viral components, which could lead to ER overload, or increase the risk of an elevated viral DNA integration in the host genome or enhanced viral cccDNA reservoir. Further studies are warranted to understand how the scFv molecules may lead to intracellular elimination of HBsAg/virion particles. Moreover, there was no evidence for a strong effect on liver integrity and functionality as reflected by the almost unchanged histology and serum liver enzyme levels, arguing against a scenario resembling a fulminant hepatitis characterised by massive destruction of liver tissue leading to liver failure.

In essence, the authors describe a novel facet of CARs. Instead of the more established function of CARs ectopically expressed in T-cells to kill target cells, an unprecedented function of CARs with a direct antiviral function by interference with viral morphogenesis is described. It will be interesting to see if this approach works for other viruses as well and how this approach can be combined with the ‘classic’ ectopic expression of CARs in T cells.